After launch, track five things: activation, meaning the share of sign-ups who complete the core journey; retention, meaning the share still active after a week and a month; the core action count per active user; the cost to acquire a user against the value of one; and where users came from. Ignore downloads, page views, total sign-ups and social followers as measures of success. They measure marketing effort, not whether the product works.
Founders track what is easy to count, and what is easy to count is mostly noise. Downloads, sign-ups and followers rise with effort and say nothing about whether the product works. This post names the handful of numbers that do, explains how to read them in the first months, and shows what to set up before launch so that they exist.
The five that matter
| Metric | What it tells you | How to measure | Healthy early sign |
|---|---|---|---|
| Activation | Whether new users reach the value | Share of sign-ups who complete the core journey within a day or a week | Rising as you fix onboarding. Absolute levels vary widely by product. |
| Retention | Whether the product is worth coming back to | Share of users active at day 7 and day 30 after sign-up, by weekly cohort | A curve that flattens rather than falling to zero |
| Core action frequency | How much value active users get | Core actions per active user per week or month | Stable or rising over time |
| Acquisition cost versus user value | Whether growth can ever pay for itself | What you spent to get a user, against what a user pays or is worth over time | A credible path to value exceeding cost, even if not yet |
| Source | Where the users who stay came from | Attribution from links, store data and simply asking | One or two sources producing users who retain |
Activation: do they reach the value?
Define the core journey precisely: for a booking app, a completed booking; for a marketplace, a first transaction; for a tool, the first result produced. Activation is the share of new sign-ups who get there. A low number means the product or the onboarding is losing people before they experience what it does. It is the first number to work on after launch, because nothing downstream can be judged until users actually reach the value.
Retention: do they come back?
The single most important measure of a new product. Group users by the week they signed up, and track what share is still active one week and one month later. A retention curve that falls and then flattens means some users have found lasting value. One that falls to zero means nobody has, and no amount of marketing will fix it. Healthy levels vary enormously by product type, so compare your own cohorts over time rather than against a benchmark. What you want to see is each week's cohort retaining a little better than the last, as the product improves.
Core action frequency: how much value?
Among users who stay, how often do they do the thing the product is for? A booking app where active users book once a month and a messaging app where they message daily are both healthy, at their own frequencies. Falling frequency among retained users is an early warning that competes with retention for attention.
Acquisition cost versus value: can this be a business?
In the first months the numbers are rough. Spend on acquisition, including your own time, divided by users acquired, against what a user pays or the value they generate. You are not looking for profitability. You are looking for a credible route to value exceeding cost as the product and the channels improve. If there is none, the pricing or the channel needs to change before the product scales.
Source: where do the good users come from?
Track where users come from and cross it with retention. The source that brings the most sign-ups is often not the one that brings users who stay. When you have money to spend on growth, this is the number that tells you where.
The vanity metrics
- Downloads and total sign-ups. Cumulative and always rising. They measure effort, not outcome.
- Page views and sessions. Traffic without the context of what people did.
- Social followers and likes. Related to marketing reach, unrelated to product value.
- Time in app. Ambiguous. A user spending ten minutes may be delighted or lost.
- Feature usage counts without a denominator. "Five hundred searches" means nothing without knowing how many users and how many searches led anywhere.
None of these are useless. They are inputs. They become dangerous when they are reported as success, which they usually are, because they always go up.
What to set up before launch
- Write down the core journey and the event that marks its completion.
- Instrument that event, sign-up, and the two or three steps between, in your analytics tool.
- Capture the source of each sign-up from links, store data or a one-question form.
- Send a test user through and confirm every event arrives.
- Build one simple view: weekly cohorts with activation, day 7 and day 30 retention.
This is a few days of work when planned, and unrecoverable when forgotten, because the data from the first weeks cannot be collected later. Our post on launching an app puts it in the pre-launch checklist.
How to read the numbers in the first month
Look daily but decide weekly. Early cohorts are small and noisy. Watch activation first, fix the biggest drop-off, then watch whether the next cohort retains better. Talk to users behind the numbers: the ten who stayed and the ten who left. The metrics tell you where to look. The conversations tell you why. Our post on why apps fail after launch puts this into a ninety-day plan.
How 7L sets this up
Instrumenting the core journey and building the cohort view is part of every first release we deliver, and we read the first weeks of numbers with the founder at the weekly review. Investors will ask for retention by cohort at the next round, and it is far easier to have it from day one than to reconstruct it. Ask us about analytics in the launch plan.
Frequently asked questions
What retention rate is good?
It depends entirely on the product. Daily-use consumer apps and monthly-use business tools have completely different healthy curves. Compare your own cohorts week over week and look for a curve that flattens rather than hitting a benchmark.
Do I need an expensive analytics tool?
No. Several capable tools are free at early-stage volumes. What matters is instrumenting the right events before launch and looking at cohorts, which any of them can do.
How many users do I need before the numbers mean anything?
Trends become readable with a few hundred users across several weekly cohorts. Before that, treat the numbers as a guide to which users to talk to, not as conclusions.
Should I share these metrics with investors?
Yes, especially retention by cohort. Investors trust founders who know their numbers and show the honest curve. A flat-lining retention curve with a plan to fix it is more credible than a rising download chart.
What is the one number if I can only track one?
Retention at day 30, by cohort. It captures whether people reached value and whether it lasted. Everything else explains why it is what it is.